far down the lowest letter of the entire font reaches below the
baseline (this is often j). C<$pos_width> is how wide the string from
from the starting position is. The total width of the string is
-C<$pos_width-$neg_width>. C<$descent> and C<$ascent> are the as
+C<$pos_width-$neg_width>. C<$descent> and C<$ascent> are the same as
<$global_descent> and <$global_ascent> except that they are only for
the characters that appear in the string. Obviously we can stuff all
the results into an array just as well: